How to Create NumPy Arrays

There are two common approaches for creating NumPy arrays:

a. Using the numpy.array function
b. Using array creation routines

Let us learn both approaches one by one.

a. numpy.array function:
The basic ndarray is created using array function in NumPy as follows –

numpy.array

The numpy.array function is a powerful constructor in NumPy, allowing

the creation of arrays with various parameters. Below is an overview of
its parameters:

numpy.array(object, dtype = None, copy = True, order = None,

subok = False, ndmin = 0)

S. No. Parameter Description

An array is returned by any object that exposes the

Object
1. array interface method, or alternatively, any
(nested) sequence.

2. Dtype Desired data type of the array, optional.

Optional specification of the desired data type for

3. Copy
the array.

Specify either C (row major), F (column major), or A

4. Order
(any) to determine the array layout.

By default, the array that is returned is coerced into

Subok
5. being a base class array. If set to true, sub-classes
will be allowed to pass through without coercion.

Defines the minimum dimensions of the resulting

6. ndmin
array.

1|Page
Let us make Python programs to understand the ndarray better.

1. Creating 1-D Array: Python program to make a one-dimensional array.

import numpy as np
a = np.array([4,6,8])
print(a)

Output
[4 6 8]

This code creates a NumPy array named a, with the elements 4, 6,

and 8, and then prints the array to the console.

2. Creating 2-D Array: Python program to create two-dimensional array.

# multi- dimension arrays

import numpy as np
a = np.array([[3, 4], [9, 6]])
print(a)

Output
[[3 4]
[9 6]]

This code creates a 2-dimensional NumPy array named a with elements

arranged in two rows and two columns and then prints the array to the
console.

3. Creating Array with minimum dimension: You can create an array with a
minimum dimension using the ndmin parameter when creating the array.
This parameter specifies the minimum number of dimensions that the
resulting array should have.

# minimum dimensions
import numpy as np
a = np.array([1, 2, 3,4,5], ndmin = 2)
print(a)

Output
[[1 2 3 4 5]]

This code uses array function to create a 2-dimensional array with

elements 1, 2, 3, 4, and 5. The ndmin=2 parameter ensures that the array

2|Page
 has a minimum of two dimensions, even though the input is a 1-
dimensional list. The output is a 2D array with a single row and five
columns.

4. Creating Array of Complex Numbers: You can create an array of

complex numbers by specifying the dtype parameter as complex
when creating the array.

# dtype parameter
import numpy as np
a = np.array([1, 2, 3], dtype = complex)
print(a)

Output
[1.+0.j 2.+0.j 3.+0.j]

The dtype parameter specifies the data type of array elements. Here,
dtype=complex is used to create an array of complex numbers from
the integer values 1, 2, and 3.

A new ndarray object can be generated using one of the array

creation routines or through the low-level ndarray constructor. Let us
learn about these array creation routines.

b. NumPy - Array Creation Routines

NumPy Array Creation Routines refer to a set of functions provided by the
NumPy library that allow users to create arrays with specific characteristics,
shapes, and initial values.
The routines are the following:

1. numpy.empty
2. numpy.zeros
3. numpy.ones
4. numpy.asarray
5. numpy.arange
6. numpy.linspace
7. numpy.logspace

3|Page
Let us discuss these routines one by one.

1. numpy.empty: It is a routine that creates an uninitialized array with a

specified shape and data type. The constructor parameters include
shape, dtype (default is float), and order (default is 'C').

Syntax:

numpy.empty(shape, dtype = float, order = 'C')

The constructor takes the following parameters.

Sr.No. Parameter & Description

Shape
1
Shape of an empty array in int or tuple of int
Dtype
2
Desired output data type. Optional
Order
3 'C' for C-style row-major array, 'F' for FORTRAN style
column-major array

Example Code:

import numpy as np
x = np.empty([3,2], dtype = int)
print(x)

Output
[[0 0]
[0 0]
[0 0]]

The numpy.empty function creates an array with uninitialized values.

This code is creating a 3x2 array with integer dtype.

2. numpy.zeros: numpy.zeros produces a new array with the given size,

where all elements are set to zero.
Syntax:

numpy.zeros(shape, dtype = float, order = 'C')

4|Page
 The constructor takes the following parameters:

Sr.No. Parameter & Description

1 Shape - Shape of an empty array in int or sequence of int

2 Dtype - Desired output data type. (Optional)

3 Order - 'C' for C-style row-major array, 'F' for FORTRAN style
column-major array

Example Code 1:

# array of three zeros. The default dtype is float

import numpy as np
x = np.zeros(3)
print(x)

Output
[0. 0. 0.]

This code creates a NumPy array named x with three elements, all
initialized to zero, and the default data type is float.

Example Code 2:

# custom type
import numpy as np
arr = np.zeros((2,2), dtype = [('a', 'i4'), ('b', 'i4')])
print(arr)

Output
[[(0, 0) (0, 0)]
[(0, 0) (0, 0)]]

The code creates a 2x2 NumPy array with a custom data type, where
each element is a tuple with two integer fields ('a' and 'b'). The array
is initialized with zeros, and the output shows the array with the
specified data type.

3. numpy.ones: It makes a new array with the specified size and data
type, where all elements are set to one.

5|Page
Syntax:

numpy.ones(shape, dtype = None, order = 'C')

The constructor takes the following parameters.

Sr.No. Parameter & Description

Shape
1
Shape of an empty array in int or tuple of int
Dtype
2
Desired output data type. Optional
Order
3 'C' for C-style row-major array, 'F' for FORTRAN style column-
major array

Example Code 1:

# Create an array of five ones. The Default dtype is float

import numpy as np
arr = np.ones(5)
print(arr)

Output
[1. 1. 1. 1. 1.]

The code uses NumPy to create an array of five floating-point elements,

all set to 1. The np.ones(5) function is used, and the resulting array is
printed.

Example Code 2:

import numpy as np
arr = np.ones([2,2], dtype = int)
print(arr)

Output
[[1 1]
[1 1]]

6|Page
 The code is creating a variable arr. It is a NumPy array with
dimensions 2x2, and all the elements in the array are set to the value
1 using np.ones() function.

Let us learn to create NumPy arrays from existing data.

4. numpy.asarray: This function is used to convert input data, such as

lists, tuples, or other array-like objects, into a NumPy array.

Syntax:

numpy.asarray(a, dtype = None, order = None)

The constructor accepts the following parameters:

Parameter
S.No. Parameter Description
Name
Input data in various forms, including lists,
1. a lists of tuples, tuples, tuple of tuples, or tuple
of lists.
By default, the data type of the input data is
2. dtype
assigned to the resulting ndarray.

C (row major) or F (column major). The

3. order
default is C.

Example Code 1:

# convert the list to ndarray

import numpy as np

x = [4,5,6]
arr = np.asarray(x)
print(arr)

Output
[4 5 6]

The Python list named x is converted into a NumPy array named arr.
The np.asarray() function is employed for this conversion.

7|Page
 Example Code 2:

# Making ndarray from a tuple

import numpy as np

x = (1,2,3)
arr = np.asarray(x)
print(arr)

Output
[1 2 3]

The code is converting a tuple named x into a NumPy array named

arr using the as array() function.

Now that you know how to create a ndarray from existing data. Let us
learn how we can create an array from numerical ranges.

5. numpy.arange (ndarray from Numerical Ranges): This function

generates an ndarray object with evenly spaced values within a
specified range.

Syntax:

numpy.arange(start, stop, step, dtype)

The constructor accepts the following parameters.

• start: The beginning of the interval. If not provided, it defaults to 0.
• stop: The end of the interval (exclusive).
• step: The spacing between values, with a default of 1.
• dtype: The data type of the resulting ndarray. If unspecified, the
data type of the input is utilized.

Example Code 1:

import numpy as np
arr = np.arange(3)
print(arr)

Output
[0 1 2]

8|Page
 The code creates an array (arr) with elements ranging from 0 to 2
using the arange function.

Example Code 2:

import numpy as np
# dtype set
arr= np.arange(3, dtype = float)
print(arr)

Output
[0. 1. 2.]

A NumPy array 'arr' is created with values 0.0, 1.0, 2.0, having a data
type of float. It's generated using arrange function.

Example Code 3:

# start and stop parameters set

import numpy as np
arr = np.arange(30,40,2)
print (arr)

Output
[30 32 34 36 38]

NumPy array 'arr' is created with values from 30 to 40 (exclusive),

incrementing by 2. It's printed using arange.

6. numpy.linspace: This function bears similarity to the arange()

function. Rather than specifying a step size, this function defines the
number of evenly spaced values within the given interval.

Syntax:

numpy.linspace(start, stop, num, endpoint, retstep, dtype)

9|Page
 The constructor accepts the following parameters.

Sr.No. Parameter Description

1. start The initial value of the sequence.

The final value of the sequence, included in

2. stop
the sequence if the endpoint is set to true.
The quantity of evenly spaced samples to
3. num
generate. The default is 50

True by default, meaning the stop value is

4. endpoint encompassed in the sequence. If false, it is
excluded.

If true, it provides samples and the step

5. retstep
between consecutive numbers.

6. dtype The data type of the output ndarray.

Example Code:

import numpy as np
arr= np.linspace(20,40,5)
print(arr)

Output
[20. 25. 30. 35. 40.]

The code uses NumPy to create an array (arr) containing 5 evenly

spaced values between 20 and 40 (inclusive). It then prints the
ndarray.

7. numpy.logspace: The function yields an ndarray comprising

numbers evenly distributed on a logarithmic scale. The start and
stop endpoints on the scale are indices of the base, typically set to 10.

Syntax:

numpy.logspace(start, stop, num, endpoint, base, dtype)

10 | P a g e
The output of the logspace function is influenced by the following
parameters.

S.No. Parameter Description

The initiation point of the sequence is denoted
1. start
as "basestart."
The concluding value of the sequence is
2. stop
referred to as "basestop."
The quantity of values within the specified
3. num
range. The default value is set to 50.
If set to true, the stop value becomes the
4. endpoint
ultimate value in the range.
The fundamental value of the logarithmic
5. base
space; by default, it is set to 10.
The data type of the resulting array. If not
6. dtype explicitly provided, it is determined by other
input parameters.

Example Code:
# set base of log space to 2
import numpy as np
arr = np.logspace(1,10,num = 10, base = 2)
print(arr)

Output
[ 2. 4. 8. 16. 32. 64. 128. 256. 512.
1024.]

This code creates an array (arr) with ten logarithmically spaced

values between 2 and 1024. The logarithmic scale has a base of 2,
achieved by setting the base parameter in the logspace function.

You now know how to create NumPy arrays. Let us know

characteristics of NumPy arrays, including details such as data type,
size, memory usage, etc.

11 | P a g e
ndArray Attributes

Attributes of ndarrays are properties or characteristics associated with

these arrays that provide information about their structure and content.
These attributes offer insights into the array's dimensions, data type,
size, and memory usage etc.
Following are the main attributes of ndArrays:

1. ndarray.shape
2. ndarray.ndim
3. numpy.itemsize
4. numpy.flags

1. ndarray.shape: This array property provides a tuple containing the

dimensions of the array and can also be utilized for resizing the array.

Example Code 1:

import numpy as np
a = np.array([[4,5,6],[7,8,9]])
print(a.shape)

Output
(2,3)

The code creates a 2x3 array named a, and prints its shape, which
is (2, 3), indicating 2 rows and 3 columns.

Example Code 2:

# this resizes the ndarray

import numpy as np

a = np.array([[4,5,6],[7,8,9]])
a.shape = (3,2)
print(a)

Output
[[4 5]
[6 7]
[8 9]]

12 | P a g e
 This code snippet creates a 2x3 array a, then resizes it to a 3x2 array
using the shape property. Finally, it prints the resized array.

NumPy offers a reshape function for resizing arrays as well.

Example Code:

import numpy as np
a = np.array([[4,5,6],[7,8,9]])
b = a.reshape(3,2)
print(b)

Output
[[4 5]
[6 7]
[8 9]]

The code creates a NumPy array a with shape (2,3), then reshapes it to
a new array b with shape (3,2), and finally prints the reshaped array b.

2. ndarray.ndim: This array attributes provides the number of

dimensions (axes) in the array.

Example Code:

import numpy as np

# Create a 1-dimensional array

arr_1d = np.array([1, 2, 3])
print(arr_1d.ndim)

# Create a 3-dimensional array

arr_3d = np.array([[[1, 2], [3, 4]], [[5, 6], [7, 8]]])
print(arr_3d.ndim)

Output
1
3

In the code arr_1d is a 1-dimensional array, and arr_3d is a 3-

dimensional array. The ndim attribute is used to query the number
of dimensions of each array.

13 | P a g e
3. numpy.itemsize: This array attributes provides the size of each
element in the array, measured in bytes.

# dtype of array is float32 (4 bytes)

import numpy as np
x = np.array([1,2,3,4,5], dtype = np.float32)
print(x.itemsize)

Output
4

Array x contains 32-bit floating-point numbers, so each element

occupies 4 bytes.

4. numpy.flags: This function returns the current values of the

attributes of the ndarray object.

Sr.No. Attribute & Description

C_CONTIGUOUS (C) - The data is organized in a single, C-

1
style contiguous segment.

F_CONTIGUOUS (F) - The data is organized in a single,

2
Fortran-style contiguous segment.

OWNDATA (O) - The array either owns the memory it uses or

3
borrows it from another object.

WRITEABLE (W) - The data area is writable; setting to False

4
locks the data, making it read-only.

ALIGNED (A) - Both data and elements are appropriately

5
aligned for the hardware.

UPDATEIFCOPY (U) - This array serves as a copy of another

6 array; deallocation updates the base array with its
contents.

14 | P a g e
Example Code:

import numpy as np
x = np.array([3,4,5,6,7])
print(x.flags)

Output
C_CONTIGUOUS : True
F_CONTIGUOUS : True
OWNDATA : True
WRITEABLE : True
ALIGNED : True
WRITEBACKIFCOPY : False

The code creates a NumPy array 'x' with elements [3,4,5,6,7] and
prints its flags indicating memory layout and properties.

15 | P a g e